Cadence’s Sherman Ikemoto on Digital Twins, Power Reality and Designing the AI Factory
The Data Center Frontier ShowShow Notes
AI data centers are no longer just buildings full of racks. They are tightly coupled systems where power, cooling, IT, and operations all depend on each other, and where bad assumptions get expensive fast. On the latest episode of The Data Center Frontier Show, Editor-in-Chief Matt Vincent talks with Sherman Ikemoto of Cadence about what it now takes to design an “AI factory” that actually works. Ikemoto explains that data center design has always been fragmented. Servers, cooling, and power are designed by different suppliers, and only at the end does the operator try to integrate everything into one system. That final integration phase has long relied on basic tools and rules of thumb, which is risky in today’s GPU-dense world. Cadence is addressing this with what it calls “DC elements”: digitally validated building blocks that represent real systems, such as NVIDIA’s DGX SuperPOD with GB200 GPUs. These are not just drawings; they model how systems really behave in terms of power, heat, airflow, and liquid cooling. Operators can assemble these elements in a digital twin and see how an AI factory will actually perform before it is built. A key shift is designing directly to service-level agreements. Traditional uncertainty forced engineers to add large safety margins, driving up cost and wasting power. With more accurate simulation, designers can shrink those margins while still hitting uptime and performance targets, critical as rack densities move from 10–20 kW to 50–100 kW and beyond. Cadence validates its digital elements using a star system. The highest level, five stars, requires deep validation and supplier sign-off. The GB200 DGX SuperPOD model reached that level through close collaboration with NVIDIA. Ikemoto says the biggest bottleneck in AI data center buildouts is not just utilities or equipment; it is knowledge. The industry is moving too fast for old design habits. Physical prototyping is slow and expensive, so virtual prototyping through simulation is becoming essential, much like in aerospace and automotive design. Cadence’s Reality Digital Twin platform uses a custom CFD engine built specifically for data centers, capable of modeling both air and liquid cooling and how they interact. It supports “extreme co-design,” where power, cooling, IT layout, and operations are designed together rather than in silos. Integration with NVIDIA Omniverse is aimed at letting multiple design tools share data and catch conflicts early. Digital twins also extend beyond commissioning. Many operators now use them in live operations, connected to monitoring systems. They test upgrades, maintenance, and layout changes in the twin before touching the real facility. Over time, the digital twin becomes the operating platform for the data center. Running real AI and machine-learning workloads through these models reveals surprises. Some applications create short, sharp power spikes in specific areas. To be safe, facilities often over-provision power by 20–30%, leaving valuable capacity unused most of the time. By linking application behavior to hardware and facility power systems, simulation can reduce that waste, crucial in an era where power is the main bottleneck. The episode also looks at Cadence’s new billion-cycle power analysis tools, which allow massive chip designs to be profiled with near-real accuracy, feeding better system- and facility-level models. Cadence and NVIDIA have worked together for decades at the chip level. Now that collaboration has expanded to servers, racks, and entire AI factories. As Ikemoto puts it, the data center is the ultimate system—where everything finally comes together—and it now needs to be designed with the same rigor as the silicon inside it.
More from The Data Center Frontier Show
In general pros and bullet points, how would you summarize what’s important in this podcast. This podcast will explore why air cooling continues to matter as liquid cooling becomes more common in AI and high-density data center environments. The goal is not to position air cooling and liquid cooling as competing technologies, but to explain how they increasingly work together as part of a broader thermal management strategy. Liquid cooling is receiving significant attention because of the growth of AI workloads and higher rack densities. It can be highly effective at removing heat directly from high-performance processors, especially in applications where traditional air cooling may no longer be sufficient on its own. However, liquid cooling does not necessarily eliminate the need for airflow elsewhere in the facility. Important points for the conversation include:
Data Center Frontier Show Podcast | DCF Trends Summit 2026 As AI infrastructure expands at unprecedented speed, data center operators face a growing financial and operational challenge: What happens to GPU systems when their first production assignment ends? In this episode of the Data Center Frontier Show, recorded onsite at DCF Trends Summit 2026 in Northern Virginia, DCF Contributing Editor Doug Black speaks with Linda Li, Chief Strategy Officer of Re-Teck / Li Tong Group, about why GPU lifecycle management must begin at the design and procurement stage—not when equipment is ready for retirement. Drawing on decades of reverse supply chain experience in telecommunications and electronics, Li explains how rapidly evolving accelerators, increasing power densities, and shifting cooling requirements are changing the economics of AI infrastructure. The conversation examines four considerations for GPU disposition: remaining operational capability, component-level economics, data security and regulatory compliance, and critical-material recovery. GPUs no longer suited to frontier AI training may retain substantial value for inference, enterprise applications, academic research, or component harvesting. Li also discusses the importance of continuous asset evaluation, secondary-market timing, secure chains of custody, and designing hardware for second and third operational lives. The central warning: Companies that fail to establish lifecycle strategies at the executive level risk creating significant amounts of stranded capital as valuable equipment sits idle while market conditions change. For AI infrastructure operators investing billions in successive generations of compute, the next life of the hardware is becoming as important to financial planning as its first deployment. Listen to the conversation for a practical examination of how GPU lifecycle planning, circular infrastructure design, and reverse supply chain management are becoming essential disciplines in the AI data center economy.
The conversation around data center power has focused on replacing diesel. But for many hyperscalers and developers, modern standby diesel remains a practical solution amid grid constraints, permitting challenges, and growing reliability demands. In this discussion, Matt Vincent, Editor in Chief of Data Center Frontier, sits down with Brad Meissner, Director of Product Management at Generac, to explore why diesel continues to play a critical role in data center power strategies—and how low-NOx engines, HVO compatibility, and next-generation emissions controls are redefining diesel technology in 2026.
AI data centers are creating a new kind of power problem. As thousands of GPUs ramp together, fast and synchronized changes in electrical demand can place new pressure on UPS systems, generators and the broader power infrastructure supporting the compute load. In this episode of the Data Center Frontier Show, DCF Editor in Chief Matt Vincent speaks with Frank DeLattre, President Americas at Musashi Energy Solutions, about how hybrid supercapacitors could help data centers manage those rapid power transients — and why AI may be forcing the industry to rethink the role of energy storage altogether. DeLattre explains the difference between storing energy and managing power, and why high-power, short-duration storage may become increasingly important as AI infrastructure scales. The discussion explores placing energy storage at multiple points in the power chain, from the UPS to systems closer to the server rack, where transients can potentially be addressed before they propagate upstream. The conversation also examines whether short-duration storage can help operators get more compute from existing electrical capacity by buffering momentary peaks, as well as the tradeoffs between energy buffering and throttling GPU performance. DeLattre also raises a provocative question around traditional UPS design: If backup generators can accept load within seconds, how much stored energy does a modern AI data center actually need for ride-through? The discussion looks ahead to a future in which operators may think less in terms of a single “UPS battery” and more in terms of an energy storage architecture — combining different technologies, durations and locations to solve different power problems across the AI data center.
AI and high-performance computing are changing how data centers are designed and powered. Rack densities have more than doubled in recent years, and power demands continue to climb. For data center operators and engineers, the pressure to scale faster, reconfigure more often, and support higher loads has made the power distribution decisions more important than ever.This QuickChat video will focus on overhead busway as the optimal power distribution system for modern data center environments. Chris Osian, Product Manager at Starline, will discuss why busway outperforms traditional RPP-based approaches as power densities increase, and what the right power distribution infrastructure needs to deliver: the capacity to support increasing power densities, the flexibility to change without downtime, and the ability to scale with the facility over time. With over 30 years of proven performance in mission-critical environments, Starline is the name data centers have trusted to deliver on all three.